1  

Trading Effects of FIFA World Cup1

Veysel Avsar2 Umut Unal3

Abstract

This study analyzes the trading effects of FIFA World Cup in two dimensions. First, we show that participating in the World Cup significantly increases exports from the participant countries to the host countries, relative to a control group of non-participants. Second, we demonstrate that trade is reasonably higher for host-participant pairs compared to other country pairs. We also provide dynamic estimates for both cases and offer plausible arguments and important channels for our findings.

JEL Classification: F19, L83

Keywords: Trade, World Cup, mega sports events

                                                                                                                         1 For helpful discussions, we thank: Kishore Dhavala, Sheng Guo, Cem Karayalcin, Dmitry Krichevskiy, Mohamed Nasseredine and Mehmet Ulubasoglu. An earlier version of this paper was circulated under the name: ‘’Trading Effects of the National Teams’ Showcase’’. 2 Corresponding author, Department of Economics, Antalya International University, Antalya, Turkey. E-mail: veysel.avsar@antalya.edu.tr. 3 Department of Economics, Turgut Ozal University, Ankara, Turkey. E-mail: uunal@turgutozal.edu.tr

2  

I. INTRODUCTION  

At the 2004 conference of the joint Turkish and Korean business council, Park Won-

jin, the chairman of the Korean-Turkish Business Council, stressed the importance of the

2002 World Cup in Korea on the swelling of bilateral trade between the two countries. It was

Turkey’s first World Cup appearance since 1954 and they won the bronze medal after an

entertaining victory over Korea. Shortly thereafter, bilateral trade between these countries

leapt to a sum of one billion dollars, which represents a 20-percent increase from the previous

year. There was another 32-percent increase, with bilateral trade reaching more than 1.3

billion dollars in 2004. Is the argument raised by Park Won-jin historically relevant? Does the

World Cup promote trade between the host and participant countries? To our knowledge, this

is the first paper to address this question empirically.

The World Cup is the most widely followed mega-sport event in the world. Unlike the

Olympic Games, national teams face a tough competition in order to qualify for the

tournament, which produces a single winner in a single sport. As opposed to the one-month

period of the actual tournament, the qualification stage takes two years and a total of

approximately 200 nations compete in around 800 games to qualify for the 32 spots in the

tournament.4 Although much lower than the cost of hosting the World Cup, national teams

still have to spend significant amounts of money to qualify for the tournament. Costs include

hiring a good manager, scouting for alternative players, and improving domestic sports

facilities for the preliminary games. Countries do not want to miss the world’s soccer feast

and wait another four years because being in the competition brings a number of benefits to

the country. In this paper, we investigate one dimension of these benefits, an improved

trading relationship between the host and participant countries.

Our empirical results suggest that the World Cup significantly increases the exports to

the host countries from the participant countries relative to non-participants and that the host-

participant pairs have higher trade with each other compared to other country pairs.  

                                                                                                                         4 Baade and Matheson (2004).

3  

The first plausible argument for our findings is the increase in visibility and exposure.

Mega sporting events bring important advertising, which increases the ability of both the host

and the participant countries to influence each other’s consumers. One consequence of this

influence is that it improves the awareness of their products, allowing these countries easier

access to each other’s markets.5

Secondly, the World Cup attracts significant numbers of tourists from the participant

countries to the host country. Although the main purpose of these travels is to attend the

games, it can also work as an important mechanism to start, build, and improve trading

networks for businessmen. In addition, potential visitors might be attracted to the venues after

the tournament due to their exposure during the games.6 As noted in Poole (2010),

international travel helps buyers and sellers transfer information about local culture, customs,

and markets, creating an efficient business relationship.7 Therefore, the World Cup can help

the host and participant countries to attenuate the informational and cultural barriers to trade.

Although the channels described above can be defined as the major mechanisms that

explain the trading effects of the World Cup, one can also note other arguments. For instance,

Rose and Spiegel (2011) point out that the hosting of large sporting events can work as a

signal for trade liberalization and increase countries’ openness significantly. Following the

same line of argument, the World Cup can also contribute to the negotiations of trade

liberalization between the host and participant countries.8

In addition, foreign business organizations in both countries may have easier access to

policy makers to lobby for lower protection and reduced bureaucracy and procedures because

of the World Cup. Therefore, larger sports events may also work as an important channel to

decrease the bilateral trading costs and thus contribute to the trade between the host and the

participant countries.

                                                                                                                         5 See Preuss (2004). 6 See Rose and Spiegel (2011). 7 There is a growing literature about the effect of international travels on trade. [Including, but not limited to, Kulendran and Wilson (2000), Aradhyula and Tronstad (2003), Poole (2010), Christea (2011)]. 8 Two examples in this category are the formation of North American Free Trade Agreement (NAFTA), which coincides with the 1994 World Cup, held in the U.S., and Turkey’s entry to the European Free Trade area, which coincides with the 1996 European Championship (Regional version of World Cup) in England.

4  

This study broadly fits into the literature that examines the effect of mega sports

events on economic activity.9 Generally, the primary focus in this literature is the cost and

benefits associated with hosting mega sports events. For instance, employing an extended

difference in difference estimation technique and using the data for largest 75 urban districts

in Germany, Hagn and Meannig (2008) have demonstrated that the 2006 World Cup has

almost no effect on the unemployment in the match venues, which is in line with the results of

a number of works, particularly those of Baade and Matheson (2000 and 2001) and Carlino

and Coulson (2004). Among these, Baade and Matheson (2001) provide estimates for the

economic impacts of Major League Baseball (MLB) by examining the data for the 23 All-Star

games during the period of 1973 through 1997. Their analysis reveals that All-Star Game

cities had employment growth below that which would have been expected.

Similarly, Baade and Matheson (2004) identify a model for the cities in the US that

are hosting the 1994 World Cup. For this purpose, income growth rates of each city for each

year between 1970 and 2000 are estimated. Thereafter, the difference between actual and

predicted growth rates is compared. The results, however, indicate an overall negative impact

on the average host city and the US economy overall. Moreover, the findings of this study of

soccer World Cup is in agreement with the results of Coates and Humphreys (1999 and

2002).

On the other hand, Lee and Taylor (2005) show that the 2002 World Cup contributed

major economic benefits to the host country: South Korea. Using an input-output framework,

they calculate that the World Cup generated an economic impact of $1.35 billion of output,

$307 million of income and $713 million of value added for South Korea.

The literature which indicates significant positive impacts of sport facilities and sports

events is, of course, not only limited to Lee and Taylor’s study. According to Plessis and                                                                                                                          9  There is also a substantial body of literature devoted to the non-market impacts of these mega sports events. For instance, Torgler (2004) reports empirical evidence of team and referee performances in the FIFA World Cup 2002. His results suggest that being a hosting nation has a significant impact on the probability of winning a game. In addition, Coupe (2007) investigates the determinants and effects of bonus schemes used during the World Cup 2006. He points out that there is not any clear indication that bonuses affected the competition outcome or the quality of the competition.

5  

Venter (2010), the immediate impact of the World Cup 2010 on the South African economy

was around 0.1 percent of GDP. Moreover, it is noted that there are some sectors enjoyed an

unseasonably good June and were beneficiaries of the tournament. Similarly, using a dynamic

inter-industry based macroeconomic simulation model, Ahlert (2007) estimates the potential,

overall and regional economic effects on the German economy because of hosting the 2006

World Cup. His results indicate a positive overall effect on GDP, private consumption and

investment. Moreover, he points out that the increase in GDP due to the pre-event phase (the

additional sport specific investments) and/or event phase (expenditures of the World Cup

tourists) is not limited to the sport-specific sectors of economy. It is also reported that there is

an expansion in non-sport specific sectors of the whole economy even from the beginning of

2003.

Our paper most closely complements the earlier work by Rose and Spiegel (2011),

which examines the trading effects of hosting the Olympics. Their results suggest that trade is

over 20 percent higher for the host countries, and unsuccessful bids to host the Olympics also

have a similar impact on exports. For the first time in the literature, our focus is on the

relationship between the host-participant pairs, rather than the effect of the large sporting

events on the host country itself.

The remainder of the paper is organized as follows. In the next section, we describe

the data. Section 3 includes the empirical specification and the results. Section 4 provides a

brief conclusion.

II. DATA  

We obtained the trade data and the other country level variables from Rose and

Spiegel (2011).10 For the trade data that is not recorded in their dataset, we utilized the IMF

Direction of Trade Statistics.11 For the World Trade Organization (WTO) membership, we

                                                                                                                         10See http://faculty.haas.berkeley.edu/arose/RecRes.htm (We appreciate their generosity.) 11 Rose and Spiegel (2011)’s data have the export values (Xij) from country i to j. For the export values from j to i, we constructed our own data.

6  

utilized the WTO website. Finally, the data on the hosts and the participants of the World Cup

were obtained from the website of FIFA (Fédération Internationale de Football Association).

Our sample includes bilateral observations for the periods between 1950 and 2006 for

196 countries. Table A1 shows the host countries for each World Cup and the corresponding

participants.

III. EMPIRICAL  SPECIFICATION  

3.1. Exports  to  the  Host  Countries:  Participants  vs.  Non-­‐participants    

Pooled OLS Estimation

We started our empirical specification by analyzing the effect of the World Cup on the

trade flows of the participant countries to the host countries, relative to non-participants. To

do so, we pooled our bilateral trade data over World Cup host and year combinations and

employed the “gravity” model of international trade, which is extensively used to estimate the

bilateral trade flows between paired countries. The following equation is estimated by OLS12:

ln(Xijt) = β0 + β1 ln(Xijt-1) + β2 ln(popit) + β3 ln(popjt) + β4 ln(GDPpcit) + β5 ln(GDPpcjt ) + β6

Common Languageij + β7 RTAijt + β8 Common Borderij + β9 Islandsij + β10 WTOijt + β11 Areaij

+ β12 Colonyij + β13 Landlockedij + β14 ln(Dij) + θ1 Participantjt + εijt (1)

where i denotes the World Cup host, j denotes the exporter country, t is the year of the

tournament, Xijt denotes the exports from j to the host countries, pop denotes population,

GDPpc is the annual real GDP per capita. Common Language and Common Border take a

value of 1 if the country pairs share the same language and the same land border, respectively.

Area is the log of the areas of the countries. RTA is a binary indicator and is unity if the

countries have a regional trade agreement. Island is the number of island countries; WTO is

the number of WTO members; and Landlocked is the number of landlocked countries in the

                                                                                                                         12Although hosting the World Cup may be thought as endogenous, participating to the tournament is exogenous. Therefore, OLS is not biased.  

7  

pair. Colony is a binary indicator and is equal to 1 if the country pairs have colonial ties. Dij

denotes the distance between i and j. In addition, we also added the lagged value of the

dependent variable to capture the effect of past trading relationships on current trade.13

The variable of interest in (1) is Participantjt, which takes a value of 1 if the country j

is qualified to the particular World Cup held in country i. Since we pooled the data for World

Cup host and year combinations, we interpret the coefficient of this variable as the increase in

the exports of the participant countries to the host countries relative to the control group of

non-participants.

Table 1 reports the coefficient estimates obtained from equation (1).  Since the year

dummies control for the invariant host country specific variables, when we pool the data over

World Cup host and year combinations, specifications in Table 1 do not include the GDP and

the population of the host country.14 To save space, we do not discuss the coefficients of the

standard gravity controls as they are not of particular interest. Consider the first column in

Table 1; the estimate of θ1 is positive and statistically significant. In terms of the magnitude

of the effect, participating in the World Cup increases exports to the host countries by 17

percent, relative to a control group of non-participant countries.15 In columns 2 to 6, we carry

out a number of experiments which were suggested by Rose and Spiegel (2011). For instance,

one could argue that the results obtained in specification 1 might be particularly driven by the

trading effects of the World Cup in a specific region. For this consideration, we removed the

observations where country j is a Latin American Country, an African country, and a Middle

Eastern Country in specifications 2, 3 and 4, respectively. Moreover, we also dropped the

observations for the poor exporters (those with real GDP per capita less than $1000 per

annum) in specification 5. Finally, we estimated the regression only for the European

countries in specification 6 given the fact that trade is especially intensive within Europe and

                                                                                                                         13See Eichengreen & Irwin, 1997; Bun & Klaassen, 2002; De Grauwe & Skudelny, 2000; Vandenbussche & Zanardi, 2010.  14  In our pooled OLS estimation, year dummies are equivalent to World Cup host dummies.    15 We use the formula used by Kennedy (1981) to convert the coefficient of the dummy variable to its true marginal effect.

8  

they have a higher probability of being in the World Cup.16 As documented, the coefficient θ1 is always positive and significant although slightly different in magnitude. Therefore, our

estimates do not result from some small subset of the data.17

Dynamic Specification

We also estimate the effect of the World Cup on the trade flows of participant

countries to the host countries, relative to a control group of non-participants using panel data.

In particular, we employ the following specification:

ln(Xijt) = β0 + β1 ln(Xijt-1) + β2 ln(popit) + β3 ln(popjt) + β4 ln(GDPpcit) + β5 ln(GDPpcjt )+ β6

WTOijt + β7 RTAijt + θ2 Participantjt + θ3 (Participantjt x Trend) + εijt + µijt (2)

where i denotes the World Cup host, j denotes the exporter country, and µijt denotes country-

pair fixed effects. There are three differences in this specification when compared to equation

(1). First, t denotes the time period between 1950 and 2006. Second, our variable of interest

Participantjt takes on a value of 1 at or before the year of the event, and the variable Trend

indicates the time trend after the World Cup takes place. It is likely that the trade effect of the

World Cup decreases as time passes after the event.18 This dynamic effect is captured by the

coefficient θ3. Further, since country-pair fixed effects control for all time-invariant factors,

we do not include them in (2), as opposed to (1).

A serious problem in estimating (2) is the serial correlation in the export series which

would bias the OLS estimates. We therefore apply instrumental variable (IV) estimation

where the second lag of the dependent variable is used as an instrument.19 To address the

concern of weak instruments, we tested the quality of our instrument. The F-test obtained in

the first stage confirmed the validity of our instrument.20

                                                                                                                         16  The results also hold for the non-Euro zone countries.  17  While we do not report here, the same pattern of results also hold for the exports originating from the host country. The results are available in the working paper revision of the article.    18  This variable is zero before the event for the particular host-participant pairs. 19  Vandenbussche and Zanardi (2010).  20  The first stage estimates are reported in Table A2.  

9  

The estimation results obtained from equation (2) are reported in Table 2. As shown,

both θ2 and θ3 are significant with the expected signs. To gauge the economic effect, consider

specification (1) in Table 2. Countries that qualified to the World Cup finals exported around

16 percent higher to the host countries for the year of the World Cup and this effect is

decreasing 0.5 percent every year. For instance, for the 5th year after the event, the effect of

World Cup on the export flows is estimated as 13 percent approximately. Similar to the

earlier section, we also check the robustness of our results in the next specifications of Table

3. As documented, our results are insensitive to all of the robustness checks described earlier.

3.2. Trading  Effects  of  the  World  Cup:  Host-­‐participant  pairs  vs.  Other  Countries  

 

Pooled OLS Estimation

Having analyzed the participant countries’ exports to host countries relative to a

control group of non-participant countries, we turn to the trading effects of the World Cup for

the host-participant pairs relative to all other country pairs.21 Hence, we pooled the bilateral

trade data over World Cup years to estimate the following model via OLS:

ln(Xijt) = β0 + β1 ln(Xijt-1) + β2 ln(popit) + β3 ln(popjt) + β4 ln(GDPpcit) + β5 ln(GDPpcjt ) + β6

WTOijt + β7 RTAijt + θ4 Participantijt + εijt + µijt (3)

where i denotes the importer and j denotes the exporter, and t denotes the years in which there

was a World Cup (1950, 54, 58….2006). Participantijt is unity if country j participated in the

World Cup organized in country i in year t. µijt denotes country-pair fixed effects. 22  

We document the estimation results of equation (3) in Table 3. Similar to earlier

estimates, we obtain positive and significant estimates for the variable of interest.

Economically speaking, in the years of the World Cup, trade is about 18-20 percent higher for

                                                                                                                         21  Notice that other country pairs include host countries and non-participants, and participants and non-hosts.  22  As opposed to the previous pooled estimation in section 3.1, we include the country-pair fixed effects in equation (3) due to the higher number of observations that make the estimation feasible with fixed effects.  

10  

host-participant combinations when compared to other country pairs. Once again, removing

some subsets of our sample does not alter our findings.

Dynamic Specification

Our last specification is the panel estimation of the permanent effect of the World

Cup on trade flows of host-participant pairs, compared to other country pairs. The following

gravity equation quantifies the model:

ln(Xijt) = β0 + β1 ln(Xijt-1) + β2 ln(popit) + β3 ln(popjt) + β4 ln(GDPpcit) + β5 ln(GDPpcjt )+ β6

WTOijt + β7 RTAijt + θ5 Participantijt + θ6(Participantijt x Trend) εijt + µijt (4)

where i denotes the importer, j denotes the exporter, and t denotes the years from 1950 to

2006. As opposed to the previous three specifications, this one includes all country pairs and

years in the sample. In line with the previous panel estimation, Participantijt takes on a value

of 1 if the country j participated in the World Cup in country i during or prior to year t and the

Trend denotes the number of years after the World Cup. We correct the bias associated with

the serial correlation of export values by applying IV regression. The first stage results

suggest that the second lag of the dependent variable is a strong instrument for the first lag.23

Table 4 illustrates the estimation results for equation 4. Similar to the findings in

Table 2, the estimates on the coefficient θ5 and θ6 suggests that host-participant pairs had

higher trade between each other as compared to other country pairs and the trade effect of the

World Cup is decreasing over time. This effect is robust to the number of robustness checks

such as dropping some of the regions and poor countries from the sample and running the

regression only for European countries.

IV. CONCLUSION  

The FIFA World Cup attracts media and sponsorship, draws thousands of

international tourists, and provides important global showcase opportunities for countries to

                                                                                                                         23  See  Table A2.  

11  

improve their visibility and exposure. It also works as a channel for the host and the

participant countries to reduce the cultural and informational barriers between them.

Therefore, the World Cup brings many mechanisms to promote trade between the host and

the participant countries. In this paper, we build on this argument and obtain strong evidence

of a reasonable trading effect of the World Cup using bilateral trade data for 196 countries

between 1950 and 2006.

We carry out our empirical analysis in two dimensions. First, we show that

participating in the World Cup significantly increases exports from the participant countries

to the host countries, relative to a control group of non-participants. Second, we demonstrate

that trade is reasonably higher for host-participant pairs compared to other country pairs. We

also provide dynamic estimates for both cases and conclude that the trade effect of the World

Cup is decreasing over time. These findings are also important in the sense that most of the

existing studies have concentrated on the economic effects of hosting large sports events. Our

study paves the way for detailed works on large sports events and the economic, social, and

cultural relationships between the host and the participant countries.

12  

REFERENCES

Ahlert G. (2006). Hosting the FIFA World Cup Germany 2006, Journal of Convention and

Event Tourism. 8(2): 57-78.

Aradhyula S. and Tronstad R. (2003). Does Tourism Promote Cross-Border Trade? American

Journal of Agricultural Economics. 85: 569-579.

Baade R. and Matheson V. (2000). An Assessment of the Economic Impact of the American

Football Championship, the Super Bowl, on Host Communities, Reflets et Perspectives. 39(2-

3): 35-46.

Baade R. and Matheson V. (2001). Home Run on Wild Pitch? The Economic Impact of Major

League Baseball’s All Star Game, Journal of Sports Economics. 2(4): 307-26.

Baade R. and Matheson V. (2004). The Quest for the Cup: Assessing the Economic Impact of

the World Cup, Regional Studies. 38(4): 343-354.

Bun, M. and Klaassen F. (2002). The Importance of Dynamics in Panel Gravity Models of

Trade, mimeo, University of Amsterdam.

Carlino G. and Coulson N.E (2004). Compensating Differentials and The Social Benefits of

the NFL, Journal of Urban Economics. 56: 25-50.

Christea A. (2011). Buyer – Seller Relationships in International Trade: Evidence from U.S.

States’ Exports and Business Class Travel, Journal of International Economics, forthcoming.

Coates, D. and Humphreys B. R. (1999). The Growth Effects of Sports Franchises, Stadia and

Arenas. Journal of Policy Analysis and Management. 18(4): 601-624.

Coates, D. and Humphreys B. R. (2002). The Economic Impact of Postseason Play in

Professional Sports. Journal of Sports Economics. 3(3): 291-299.

13  

Coupe, T. (2007). Incentives and Bonuses – The Case of the 2006 World Cup. Kyklos. 60(3):

349-358.

De Grauwe P. and Skudelny F. (2000). The Impact of EMU on Trade Flows, Review of

World Economics, (Weltwirtschaftliches Archiv). 126: 381-402.

Eichengreen B. and Irwin D.A. (1997). The Role of History in Bilateral Trade Flows, in J.A.

Frankel (ed.), Issues in Regionalism, Chicago: University of Chicago Press.

Hagn F. and Maennig W. (2008). Employment Effects of the Football World Cup 1974 in

Germany, Labour Economics. 15: 1062-1075.

Kennedy P. (1981). Estimation with Correctly Interpreted Dummy Variables in

Semilogarithmic Equations, American Economic Review. 71: 801.

Kulendran N. and Wilson K. (2000). Is There a Relationship Between International Trade and

International Travel? Applied Economics. 32: 1001-1009.

Lee C. and Taylor T. (2005). Critical Reflections on the Economic Impact Assessment of a

Mega Event: The Case of 2002 FIFA World Cup. 26: 595-603.

Plessis S. and Venter C. (2010). The Home Team Scores! A First Assessment of the

Economic Impact of World Cup 2010, Stellenbosch Economic Working Papers: 21/10

Poole J. (2010). Business Travel as an Input to International Trade, mimeo, UC Santa Cruz.

Preuss H. (2010). The Economics of Staging the Olympics, Edward Elgar, Northampton.

Rose, A. and Spiegel M. (2011). The Olympic Effect, Economic Journal. 121: 652-677.

Torgler, B. (2004). The Economics of the FIFA Football Worldcup. Kyklos. 57(2): 287-300.

Vandenbussche H. and Zanardi M. (2010). The Chilling Trade Effects of Antidumping

Proliferation, European Economic Review. 54(6): 760-777.

   

  14  

TABLE A1. FIFA WORLD CUP: 1950-2006 Year Host Participants

1950 Brazil Brazil, Italy, England, Uruguay, Sweden, Switzerland, Spain, Bolivia, Chile, Paraguay, Yugoslavia, India, Mexico, Portugal, United States, France

1954 Switzerland Switzerland, Uruguay, Brazil, Hungary, Austria, England, West Germany, Yugoslavia, France, Italy, Czechoslovakia, Turkey, Belgium, Mexico, South Korea, Scotland

1958 Sweden Sweden, West Germany, Austria, France, Czechoslovakia, Hungary, Soviet Union, Yugoslavia, England, Northern Ireland, Scotland, Wales, Argentina, Brazil, Mexico, Paraguay

1962 Chile Chile, Brazil, Argentina, Uruguay, Czechoslovakia, England, Soviet Union, West Germany, Italy, Hungary, Spain, Yugoslavia, Bulgaria, Colombia, Mexico, Switzerland

1966 England Brazil, Argentina, Chile, Uruguay, England, Italy, Portugal, West Germany, France, Hungary, Switzerland, Soviet Union, Bulgaria, North Korea, Mexico, Spain

1970 Mexico Brazil, Italy, Germany FR, Uruguay, England, Mexico, Peru, Soviet Union, Belgium, Bulgaria, Czechoslovakia, El Salvador, Israel, Morocco, Romania, Sweden

1974 Germany West Germany, Italy, Netherlands, Scotland, Bulgaria, Germany, Poland, Yugoslavia, Brazil, Argentina, Chile, Uruguay, Australia, Haiti, Sweden, Zaire

1978 Argentina Argentina, Germany, Netherlands, Brazil, Italy, Sweden, Mexico, Peru, Hungary, Poland, Scotland, Spain, Austria, France, Iran, Tunisia

1982 Spain Spain, Argentina, Brazil, Italy, Germany, England, Austria, Soviet Union, Hungary, Poland, Czechoslovakia, Yugoslavia, Belgium, Scotland, Northern Ireland, France, Chile, Peru, Algeria, Cameroon, Kuwait, El Salvador, Honduras, New Zealand

1986 Mexico Germany, France, Belgium, Brazil, England, Mexico, Spain, Bulgaria, England, Mexico, Spain, Bulgaria, Denmark, Italy, Morocco, Paraguay, Poland, Uruguay, Soviet Union, Algeria, Canada, Hungary, Iraq, Korea Republic, Northern Ireland, Portugal, Scotland

1990 Italy Italy, Argentina, Brazil, Germany, Belgium, England, Austria, Netherlands, Scotland, Spain, Soviet Union, Yugoslavia, Colombia, Czechoslovakia, Ireland, Romania, Sweden, Uruguay, Cameroon, Costa Rica, Egypt, South Korea, United Arab Emirates, United States

1994 USA United States, Germany, Argentina, Belgium, Brazil, Italy, Bulgaria, Ireland, Netherlands, Romania, Spain, Russia, Greece, Norway, Sweden, Switzerland, South Korea, Saudi Arabia, Cameroon, Morocco, Nigeria, Bolivia, Colombia, Mexico

1998 France France, Brazil, Argentina, Germany, Italy, Netherlands, Romania, Spain, Cameroon, Morocco, Nigeria, South Africa, Tunisia, Jamaica, Mexico, United States, Austria, Belgium, Bulgaria, Croatia, Denmark, England, Norway, Scotland, Yugoslavia, Chile, Colombia, Paraguay, Iran, Japan, South Korea, Saudi Arabia

2002 Korea and Japan Argentina, Brazil, Germany, Italy, France, Japan, South Korea, Spain, Belgium, Croatia, Denmark, England, Ireland, Poland, Portugal, Russia, Slovenia, Sweden, Turkey, China, Ecuador, Paraguay, Saudi Arabia, Uruguay, Cameroon, Costa Rica, Mexico, Nigeria, Senegal, South Africa, Tunisia, United States

2006 Germany Iran, Japan, Saudi Arabia, South Korea, Angola, Ivory Coast, Ghana, Togo, Tunisia, Costa Rica, Mexico, Trinidad and Tobago, United States, Argentina, Brazil, Ecuador, Paraguay, Australia, Croatia, Czech Republic, England, France, Italy, Netherlands, Poland, Portugal, Serbia, Spain, Switzerland, Sweden, Ukraine

   

  15  

TABLE A2. TESTING INSTRUMENT QUALITY

First Stage Estimates

Table 2 (1) Table 4 (1)

First lag First lag

Second lag 0.130 0.210 (187.20)*** (40.57)*** Second lag Participantj 0.241 0.146 (4.76)*** (1.02) WTO member 0.078 0.065 (4.65)*** (5.43)*** Log Populationi -0.014 0.233 (0.46) (0.84) Log Populationj -0.014 -0.009 (0.48) (0.05) Log (Real GDP p/c)i -0.085 -0.222 (4.76)*** (0.94) Log (Real GDP p/c)j -0.045 -0.075 (2.77)*** (1.50) RTA 0.010 0.073 (0.68) (0.86) Observations 63589 327537 R-squared 0.10 0.19 F stat 562.11 34.02

Notes: Robust t-statistics in parentheses. * significant at 10%; ** significant at 5%; *** significant at 1%. Year dummies and constant terms included but suppressed. Constant terms included.

   

  16  

Notes: Robust t-statistics in parentheses. * significant at 10%; ** significant at 5%; *** significant at 1%. Year dummies and constant terms included but suppressed.

Table 1. OLS Estimation Results (Pooled over World Cup Hosts & Years Combinations)

Dependent variable: Log of Exports to the host countries [ln(Xijt)]

Participants vs. Non-participants All sample (1) Drop Latin

American Countries (2)

Drop African Countries (3)

Drop Middle Eastern Countries (4)

Drop Poor Countries (5)

EU zone Countries

ln(Xijt-1)] 0.863 0.855 0.862 0.862 0.872 0.919 (45.07)*** (39.00)*** (38.17)*** (44.31)*** (41.41)*** (22.85)*** Participantjt 0.172 0.162 0.181 0.179 0.164 0.198 (2.04)** (2.09)** (2.49)** (2.85)** (2.24)** (2.34)** WTO member 0.136 0.183 0.042 0.197 0.066 0.176 (2.82)** (2.07)** (2.23)** (2.15)** (2.38)** (0.65) Log Distanceij -0.198 -0.193 -0.262 -0.176 -0.150 -0.74 (0.28) (0.23) (0.95) (0.02) (0.69) (0.48) Islands -0.274 -0.256 -0.366 -0.264 -0.171 -0.054 (2.46)** (2.13)** (2.92)*** (2.25)** (1.34) (0.15) Log Populationjt 0.449 0.434 0.457 0.446 0.469 0.413 (1.73)* (1.92)* (1.49) (1.72)* (1.71)* (1.04) Log (Real GDP p/c)j 0.365 0.370 0.317 0.370 0.382 0.049 (1.66)* (1.73)* (1.75)* (1.71)* (1.97)** (1.71)* Landlocked -0.197 -0.216 -0.244 -0.205 -0.165 -0.056 (1.80)* (1.80)* (1.95)* (1.84)* (1.47) (0.30) Common Languageij 0.202 0.192 0.203 0.229 0.289 0.270 (1.66)* (1.79)* (1.74)* (1.59) (1.86)* (1.93)* RTAij 0.085 0.191 0.011 0.118 0.025 0.052 (0.72) (1.36) (0.09) (0.95) (0.21) (0.21) Common Borderij 0.480 0.310 0.399 0.498 0.550 0.376 (2.00)** (1.89)* (1.85)* (2.04)** (2.22)** (2.41)** Log Product Land Areasij -0.011 -0.021 -0.008 -0.018 -0.001 -0.013 (0.45) (0.74) (0.28) (0.70) (0.02) (0.19) Colonyij 0.345 0.317 0.202 0.345 0.515 0.468 (1.81)* (1.66)* (0.91) (1.72)* (2.32)** (2.14)** Observations 1331 1114 997 1229 1146 770 R-squared 0.83 0.82 0.84 0.83 0.84 0.71

   

  17  

Table 2. IV Panel Estimation – Country Pair Fixed Effects

Dependent variable: Log of Exports to the host countries [ln(Xijt)]

Participants vs. Non-participants All sample Drop Latin

American Countries

Drop African Countries

Drop Middle Eastern Countries

Drop Poor Countries

EU Zone Countries

[ln(Xijt-1)] 0.749 0.722 0.731 0.711 0.732 0.722 (105.03)*** (83.41)*** (90.76)*** (86.32)*** (67.34)*** (76.44)*** Participantjt 0.158 0.153 0.154 0.149 0.141 0.148 (2.92)*** (2.11)** (3.31)*** (3.16)*** (2.99)*** (2.81)*** Participantjt x Time trend -0.005 -0.005 -0.004 -0.005 -0.005 -0.005 (2.59)*** (1.98)** (2.65)** (2.13)** (1.97)** (2.67)*** WTO memberijt 0.035 0.048 0.027 0.051 0.043 0.067 (1.24) (1.60) (0.41) (1.85)* (1.59) (1.84)* Log Populationit 0.331 0.397 0.467 0.387 0.265 0.013 (2.03)** (2.01)** (2.64)** (3.06)*** (3.27)*** (0.14) Log Populationjt 0.232 0.212 0.241 0.216 0.231 0.223 (296.)*** (4.76)*** (3.54)*** (3.87)*** (3.96)*** (1.91)* Log (Real GDP p/c)it 0.599 0.561 0.487 0.327 0.466 0.173 (2.02)** (1.99)** (2.73)** (2.43)** (2.47)** (2.45)** Log (Real GDP p/c)jt 0.410 0.467 0.432 0.327 0.376 0.061 (0.81) (0.32) (0.62) (0.39) (0.56) (1.45) RTAijt 0.032 0.061 0.018 0.043 0.025 0.038 (1.03) (1.54) (0.58) (1.91)* (1.12) (0.22) Country pair fixed effect Yes Yes Yes Yes Yes Yes R-squared 0.80 0.78 0.80 0.80 0.80 0.45 Observations 63589 53095 46999 58333 57001 38732

Notes: Robust t-statistics in parentheses. * significant at 10%; ** significant at 5%; *** significant at 1%. Year dummies and constant terms included but suppressed.

   

  18  

Table 3. OLS Estimation (Pooled over the World Cup years)

Dependent variable: [ln(Xijt)]

Host- Participant pairs vs. Others All sample Drop Latin

American Countries

Drop African Countries

Drop Middle Eastern Countries

Drop Poor Countries

EU zone Countries

[ln(Xijt-1)] 0.621 0.616 0.621 0.624 0.619 0.640 (190.87)*** (174.93)*** (171.07)*** (183.72)*** (181.74)*** (102.83)*** Participantijt 0.191 0.204 0.208 0.200 0.199 0.188 (2.05)** (2.81)*** (2.24)** (2.10)** (2.14)** (1.97)** WTO member 0.004 0.004 0.028 0.003 0.015 0.012 (0.13) (0.13) (0.81) (0.09) (0.47) (0.87) Log Populationit 0.381 0.344 0.312 0.370 0.306 0.002 (1.78)* (1.79)* (1.60) (1.76)* (1.74)* (1.77)* Log Populationjt 0.172 0.195 0.137 0.161 0.153 0.003 (1.29) (1.58) (1.74)* (0.98) (0.90) (0.82) Log (Real GDP p/c)it 0.720 0.843 0.839 0.839 0.711 0.067 (1.73)* (1.84)* (2.06)** (1.18) (1.20) (1.79)* Log (Real GDP p/c)jt 0.637 0.509 0.618 0.632 0.646 0.123 (2.15)** (2.29)** (2.00)** (2.87)** (2.02)** (1.51) RTAijt 0.016 0.011 0.035 0.005 0.019 0.035 (0.59) (0.36) (1.18) (0.17) (0.66) (0.57) Country pair fixed effect Yes Yes Yes Yes Yes Yes R-squared 0.39 0.39 0.39 0.39 0.39 0.42 Observations 73875 64344 58259 67532 68038 35921

Notes: Robust t-statistics in parentheses. * significant at 10%; ** significant at 5%; *** significant at 1%. Year dummies and constant terms included but suppressed.

   

  19  

Table 4. IV Panel Estimation

Dependent variable: [ln(Xijt)]

Host-Participant pairs vs. Others

All sample Drop Latin American Countries

Drop African Countries

Drop Middle Eastern Countries

Drop Poor Countries

EU Zone Countries

[ln(Xijt-1)] 0.698 0.608 0.682 0.624 0.654 0.589 (104.03)*** (106.13)*** (153.09)*** (156.48)*** (195.31)*** (117.54)*** Participantijt 0.112 0.118 0.111 0.118 0.117 0.124 (2.91)** (2.01)** (2.13)** (2.17)** (2.02)** (3.16)** Participantijt X Time trend -0.004 -0.005 -0.005 -0.005 -0.005 -0.004 (2.13)** (3.15)*** (3.41)*** (2.06)** (1.99)** (5.46)*** WTO member 0.016 0.018 0.022 0.016 0.019 0.026 (2.28)** (2.21)** (2.09)** (1.22) (1.69)* (2.08)** Log Populationit 0.035 0.026 0.027 0.024 0.027 0.027 (1.17) (0.96) (1.14) (0.37) (1.66)* (2.01)** Log Populationjt 0.181 0.214 0.318 0.317 0.272 0.258 (2.73)** (2.88)** (2.18)** (2.09)** (2.94)** (2.52)** Log (Real GDP p/c)it 0.312 0.387 0.274 0.323 0.3216 0.187 (0.45) (0.37) (0.55) (0.58) (0.65) (0.18) Log (Real GDP p/c)jt 0.602 0.476 0.420 0.478 0.553 0.422 (3.76)*** (3.77)*** (2.26)** (2.87)** (2.63)** (5.65)*** RTAijt 0.001 0.004 0.002 0.001 -0.002 0.001 (0.08) (0.60) (0.44) (0.12) (0.77) (0.97) Country pair fixed effect Yes Yes Yes Yes Yes Yes R-squared 0.83 0.82 0.82 0.83 0.78 0.76 Observations 327537 285402 257457 299657 300999 145203

Notes: Robust t-statistics in parentheses. * significant at 10%; ** significant at 5%; *** significant at 1%. Year dummies and constant terms included but suppressed. Constant terms included